Enantioselective alkylation of ketones via chiral, nonracemic

Jun 1, 1981 - Enantioselective alkylation of ketones via chiral, nonracemic lithioenamines. An asymmetric synthesis of .alpha.-alkyl and .alpha.,.alph...
0 downloads 0 Views 815KB Size
308 1

J. Am. Chem. SOC. 1981, 103, 3081-3087 spectra of the individual components. In the 'H N M R spectrum, all signals of the phenol are shifted downfield, most significantly the hydroxyl proton (Table 11). In the 13Cspectrum, the principal changes are observed for the ipso carbon. For benzophenone we observed no change in the 'HN M R spectrum but a downfield shift of the carbonyl carbon resonance. These shifts suggest a weak interaction between the hydroxyl proton and the carbonyl group in benzene-d6 solutions. Due to the close proximity of the carbonyl moiety and the phenolic hydrogen in the aggregate, it is not impossible that the photoexcited benzophenone molecules react with the phenol predominantly in the singlet state before intersystem crossing to the triplet state can occur. The resulting change in spin multiplicity ( p > 0 p < 0) would explain the observed change in signal directions [ortho, para (meta): E(A) A(E)]. The concept of singlet quenching in a complex suggests several control experiments. For example, complex formation should be sensitive to steric hinderance. This is confirmed by the system 2,6-di-tert-butylphenol-benzophenone-benzene-d6.Neither the 'H N M R nor the 13CN M R spectrum of 2,6-di-tert-butylphenol is affected by admixture of benzophenone (Table 11), and the observed polarization is that expected for triplet quenching. However, other attempts to support this concept did not produce unambigous results. Since the observed effects represent a competition between singlet and triplet state reactions, the addition of a triplet quencher such as biphenyl should enhance the effects due to the singlet reaction.ls Also the relative intensity of the CIDNP effects as a function of phenol concentration should provide information about the relative amounts of singlet and triplet reaction. The results in both experiments are ambiguous, perhaps suggesting additional interactions. All of the CIDNP results discussed thus far reflect the hyperfine coupling pattern of the phenoxyl radical. In certain experiments,

-

-

(15) (a) Wagner, P. J. J. Am. Chem. Soc. 1967,89,2820. (b) Hammond,

G. S.; Caldwell, R. A.; King, J. M.; Kristinsson, H.; Whitten, D. G. Photochem. Photobiol. 1968, 7, 695.

however, the polarization observed for the phenol does not reflect the spin densities of this intermediate. For example, the irradiation of DDBP in a CD,CN solution containing 4-tert-butylphenol led to a CIDNP spectrum which showed both ortho and meta protons in emission (Table I). The polarization of the ortho protons is that expected for a phenoxy radical, but the polarization of the meta protons appears to be inconsistent with that intermediate. In order to explain this effect, we invoke the cross-polarization mechanism previously delineated for anilinium radical cations16 and for phenoxyl radicals." This mechanism involves transfer of polarization from a strongly polarized nucleus to a nucleus having weak or negligible hyperfine interaction. In order for the polarization transfer to occur these nuclei must be coupled indirectly and must experience a periodic exchange between paramagnetic and diamagnetic states. In the case of 4-tert-butylphenol, the polarization of the ortho protons (strong hfc) is transferred to the meta protons (weak hfc) so that both are observed in emission. The occurrence of polarization transfer will depend critically on factors which influence the rate of exchange such as solution acidity. Accordingly, the cross-polarization mechanism cannot be expected to be generally operative. By applying CIDNP techniques we have been able to identify phenoxyl radicals as intermediates in the quenching of photoexcited ketones by simple phenols and have elucidated several facets of the reaction as well as the polarization mechanism. We are extending the application of these techniques to phenolic resin systems in order to determine the composition of mixed resins. Acknowledgment. The author is indebted to H. D. Roth for helpful discussions. (16) (a) Closs, G. L.; Czeropski,M. S.J. Am. Chem. Soc. 1977,99,6127. (b) Closs, G. L.; Czeropski, M. S. Chem. Phys. Left., 1978, 53, 321. (c) Hendricks, B. M. P.; Walter, R. I.; Fischer, H. J. Am. Chem. Soc. 1979,101, 2378. (17) (a) Kaptein, R., Nature (London) 1978, 274, 293. (b) Garssen, G.

J.; Kaptein, R.; Schoenmakers,J. G. G.; Hilbers, C. W. Proc. Nufl.Acud. Sci. U.S.A. 1978, 75, 5281.

Enantioselective Alkylation of Ketones via Chiral, Nonracemic Lithioenamines. An Asymmetric Synthesis of &-Alkyl and a,cu'-Dialkyl Cyclic Ketones A. I. Meyers,* Donald R. Williams,1eGary W. Erickson,lbSteven White, and Melvin Druelinger Contribution from the Department of Chemistry, Colorado State University, Fort Collins, Colorado 80523. Received October 27. 1980

Abstract: Chiral imines (S)-2 are readily prepared from cyclic ketones and the chiral methoxyamine, (S)-l. Metalation and alkylation followed by imine hydrolysis lead to 2-alkylcycloalkanones 5 in 87-100% enantiomeric purity. A method to determine % ee of 2-alkylcyclohexanonesvia their diastereomericacetals by I3C NMRis also described. Dialkylation of the chiral imines (S)-2 has produced 2,6-dialkylcyclohexanones,which for the dimethyl case were formed in 85% ee.

The alkylation of ketone enolates is among the most fundamental and widely used methods for carbon-carbon bond formation. Unfortunately, this process does not lend itself to enantioselective control due to the symmetric nature of the enolate mystem, and, thus, asymmetric C-C bond-forming reactions in simple, flexible, structures is not feasible. From Scheme I nu-

cleophilic attack by the enolate would generally be expected to introduce an electrophile above and below the mystem with equal facility and thus provide equal amounts of the enantiomeric ketones ( k , N k,; AAG* N 0). With the advent of metalloenamines2 as enolate equivalents came the opportunity to introduce a chiral environment which could influence the direction and the rate of

(1) (a) Taken in part from the Ph.D. Thesis of D.R.W. 1978; (b) National Institutes of Health Postdoctoral Fellow, 1978-1980.

(2) (a) Stork, G.; Dowd, S.J. Am. Chem. Soc. 1963,85, 2178. (b) Wittig, G.; Frommeld, H. D.; Suchanek, P. Angew. Chem. 1963, 75, 978.

~~

0002-7863/81/1503-3081$01.25/0 0 1981 American Chemical Society

3082 J . Am. Chem. SOC.,Vol. 103, No. 11, I981

Meyers et al.

Scheme I

k

L E@

Y-

ligand, the degree of asymmetric induction fell considerably short of that observed Recently, data has appearedlOJ1which indicate that other factors may also contribute to the high enantiomeric excess of the carbonyl products. These involve the use of various bases, added salts, geometry of the imines, and stereoelectronic effects in the deprotonation step. These aspects of the process and its true mechanistic pathway are still under active investigation and undoubtedly will be sorted out in the future. However, the synthetic practitioner need not await these findings in order to utilize this technique successfully. We now describe in full our synthetic results concerning a large number of chiral a-substituted, as well as a,a'-disubstituted ketones in the cyclic series. Alkylation of acyclic and macrocyclic ketones are the subject of the adjoining paper since they involve another aspect, namely, E,Z isomerization of the lithioenamine which give rise to either (S)-or (R)-ketones. From Table I it is seen that enantioselectivealkylation of a variety of cyclic ketones (via 2) have been carried out. For the cyclohexanones, the enantiomeric excess ranges from 87 to 100% and furnishes the (S)-2-alkylcyclohexanones(entries 1-3). In a similar fashion, the 2-benzyl-, 2-alkyl-, 2-acetonyl-, and 2-(2-methoxyethyl)cyclohexanones (entries 4-7) gave the (R)-enantiomer due to reversal of the Cahn-Ingold-Prelog priority assignment but possess the same sense of absolute stereochemistry as seen from their respective CD maxima. For those a-alkycycloalkanones in Table I which were not previously described, methods for determining the % ee were required. For 2-ethyl- and 2-benzylcyclohexanone their acetals 9 were prepared from (2R,3R)-(-)-2,3-butanediol in such a manner which precluded any racemization. On the basis

yo E

electrophilic entry and provide the bias necessary to form the alkylated products in disproportionate amounts (Scheme 11). This concept has been tested earlier by Horeau3 and Yamada4 with varying degrees of success leading to enantiomerically enriched a-substituted ketones in 25-45% enantiomeric excess. In 1976, a preliminary report from this laboratory' described the asymmetric alkylation of cyclohexanones via certain lithioenamines furnishing a-alkylcyclohexanones in 90-100% ee. The basis for the efficient enantioselective alkylation rested on the induced rigidity of the lithioenamine by an internal ligand (OMe)6 which was absent in the earlier work by Horeau and Yamada. Thus, starting from a readily available chiral amine 1 obtained by reduction of (S)-phenylalanine, we obtained and transformed the imine of cyclohexanone 2 into its lithioenamine 3 by using lithium

""A' Me

I

R X (-78')

qo+

*g Me

HO

Me

R

0

(3) Mea-Jacheet, D.; Horeau, A. Bull. SOC.Chim. Fr. 1968, 4571. (4) Kitomoto, M.; Hiroi, K.; Terashima, S.;Yamada, S. Chem. Pharm. Bull. 1974, 22, 459. ( 5 ) Meyers, A. I.; Williams, D. R.; Druelinger, M. J . Am. Chem. Soc.

(-1 -9 of a report by Wynberg,'* diasteromeric ketals derived from 3-substituted cyclohexanones and (-)-2,3-butanediol gave 13C N M R spectra which exhibited peak separation at the 2- and 6-carbons of 1 ppm and allowed integration leading to diastereomeric ratios. On this basis we prepared 9 from the 2-substituted cyclohexanones. It was of prime importance that no epimerization occurred during the ketal formation. Initially, we examined racemic 2-methyl- and 2-(n-propyl)cyclohexanone and found that two peaks of equal intensity appeared at -36-37 ppm (C-6) and another two peaks at -40-44 ppm (C-2) with PA6 of 0.4-0.8 ppm. In order to avoid kinetic resolution during the ketal formation, we utilized approximately 100% of excess diol. Next, to test the validity of this enantiomeric determination, we utilized 2-methyl- and 2-(n-propyl)cyclohexanone of known enantiomeric purity. The ketals were formed and, as seen from Table 11, the diastereomeric ratios from I3C N M R were in excellent agreement with the % ee via rotation data. This provided sufficient confidence that no epimerization was occurring during the ketal formation. In this manner, the enantiomeric purity of 2-ethyland 2-benzylcyclohexanone were determined, and the rotation expected for optically pure materials is given in the final column of the table. The other cyclohexanone derivatives in Table I readily succumbed to enantiomeric determinations by using chiral lanthanide-induced shift reagents and possess ee's which are within f2%. Enantiomeric purities of the cycloheptanones, cyclooctanones, tetralones, and indanones were likewise determined by LISR techniques. Unfortunately, the chiral ketals from (-)-2,3-butanediol for ketones other than cyclohexanones did not lend themselves to 13C N M R analysis since no reliable peak separations were observed. The same was found true for acyclic

1976,98, 3032. (6) The use of methoxyl for chelaton has becn demonstrated in oxazolines (Meyers, A. I. Acc. Chem. Res. 1978, 11, 375). (7) Hashimoto, S.; Koga, K. Chem. Pharm. Bull. 1979, 27, 2760. (8) Enders, D.; Eichenauer, H. Chem. Ber. 1980,112, 2933. (9) Whitesell, J. K.; Whitesell, M. A. J . Org. Chem. 1977, 42, 377.

(10) Frascr, R. R.; Akiyama, F.;Banville, J. Tetrahedron Lett. 1979,3929. (1 1) Davenport, K. G.; Eichenauer, H.; Enden, D.; Newcomb, M.; Bergbreiter, D. E. J . Am. Chem. SOC.1979, 101, 5654. (12) Heimstra, H.; Wynberg, H. Tetrahedron Lett. 1977, 2183.

(2R,3R)-

W-3

4

diisopropylamide. A key feature of 3 was the presumed formation of a rigid 5-membered chelate whose topology would influence the direction of entry by the alkyl halide, RX. Alkylation led to 4 by predominant entry from the si face (frontside) of the cyclohexenyl moiety producing, after hydrolysis, the (S)-Zalkyl ketone, 5. Recovery of the chiral methoxyamine (S)-1, for recycling, was also possible since no racemization took place during the entire sequence. A number of other successes using the "rigid lithioenamine" concept have been described7+ since the original report in 1976 and have involved the use of various chiral amines 6-8. In these instances, alkylation of the cyclohexanone imines,

-6

7

0

as well as other imines, gave a-alkyl ketones in very high enantiomeric excesses (90-10096) attesting to the rather general nature of this asymmetric synthetic method. In the absence of an internal

-

J. Am. Chem. SOC.,Vol. 103, No. 11, 1981 3083

Enantioselective Alkylation of Ketones Scheme I1

Table I. Chiral ZSubstituted Cycloalkanones 5 entry

ketone

1

cyclohexanone

Me1

+12.2 (4, MeOH)

87

(ab

+2120

2

cyclohexanone

EtI

+24.1 (4, MeOH)

94 (aC

+2200

3

cyclohexanone

n-PrI

+27.9 (4, MeOH)

99 ( a d

+2480

4

cyclohexanone

allyl Br

+15.8 (3, MeOH)

99 (R)e

+2190

5

cyclohexanone

PhCH, Br

+41.4 ( 5 , MeOH)

88 (R)'

+1750

6

cyclohexanone

+23.5 (6, MeOH)

80 ( R ) f

&&

7

cyclohexanone

8

cycloheptanone

Me1

+71.5 (6, CHC1,)

85 ( S ) f ~ 2 2

+ 28 30

9

cycloheptanone

I / V O M e

+36.4 (4, CHC1,)

81 (R)f

+2158

10

cyclooctanone

Me1

+8.07 (7, CHCl,)

20 (S)f

+572

11

1-tetralone

Me1

-40.5 (3, dioxane)

79 (s)f*g

-939h

12

1-indanone

Me1

-7.2 ( 5 , dioxane)

17 (S)fJ

+145'

+0.38 (9, CHC1,)

M -e'

90 ( R ) f

Yield based on imine 2. Beard, C.; Djerassi, C.; Sicher, J.; Sipos, F.; Tichy, M. Tetrahedron 1963,19, 919 report [ a ] D + 14". Determined by 13C NMR on the chiral dioxolane 9. Hiroi, K.; Achiwa, K.; Yamada, S.Chem. Pharm. Bull. 1972,20, 246 report [ a ] D-28.2' for R enantiomer. e Kitomoto, M.; Hiroi, K.; Terashima, S . ; Yamada, S. Chem. Pharm Bull. 1974,22, 459. f Determined by use of chlral shift reagent, tris[ (3-(heptafluoropropyl)hydroxymethylene)al-camphorato]europium(III). Jaouen, G.; Meyer, A. J. A m C h e m SOC. 1975,97, At 333 nm. I At 316 nm. 4667 report -51.2' for (S)-(-)-2-methyl-l-tetralone and -42" for (R)+)-2-methylindanone. Table 11. Determination of Enantiomeric Purity of Ketals 9 via 2-substituted cyclohexanone Me n-Pr Et PhCH

NMR 13CNMR

[,IMeoH,-, +11.4 -19.9 +24.2 +41.4

% eea 81.6 71.4 b C

ketal (% yield) 92 84 85 87

C-2 39.7 44.5 46.6 46.9

C-6 36.5 36.5 36.5 36.6

% eed,e 82 72 94 89

[a 1maxn

+14.0 -27.ge +25.7 +46.5

Based on literature values, see Table I (ref b, d ) . Helmchen-Zeier, R. E. thesis, ETH-Zurich, 1973 reports 2-ethylcyclohexanone as +19" (c 0.26, EtOH) for the @)-(+)-ketone. Determined by integration of the peaks at C-2 and C d , see Table IV for Not previously reported. complete spectral d a t a e This ketone was prepared by using (R)-1. e TI and NOE measurement showed no difference for C-2 and C d ffl each diastereomer or racemic ketals.

3084 J. Am. Chem. SOC.,Vola103, No. 11, 1981

Meyers et al.

ketones which also failed to show diastereomeric splitting of carbon peaks. As seen from Table I, all of the a-substituted ketones prepared possessed the absolute configuration in the major enantiomer consistent with the operational mechanism given earlier, namely, frontside entry of the alkyl halide on 3. This heuristic sequence may be used with some confidence in predicting the absolute configuration of a-substituted cycloalkanones formed under these reaction conditions.13 The two examples shown in Table I which resulted in rather poor enantiomeric purity, 2-methylcyclooctanone and 2methylindanone, were found to be victims of extensive racemization during the hydrolysis step to remove the imine (4 5). That the low ee's for these two examples is the result of the hydrolysis (NaOAc-HOAc, pentane-water) rather than poor selectivity in the alkylation step was shown by examining the H1 N M R spectra of the diastereomers 10 and 11 prior to hydrolysis.

LDA-RX (syn-a I ky'n)

(ss) -12 (syn-axial)

I

-

Ph Ph

SS, 9 6 ? 2 % SR,

4i2%

RX

SR,

' 4

(SS)-X

7i3%

The newly introduced methyl groups showed doublets which integrated to >95% indicating the diastereomeric ratios were indeed quite high. Furthermore, racemic 2-methylcyclooctanone and 2-methyl-1-indanone as their imines 10 and 11 gave nicely separated pairs of doublets (1:l) for the methyl group. In order to validate the accuracy of the method, we made two further checks. The % ee of 2-methylcycloheptanone was found by using LISR (Table I, footnoten, to be within 5% of that found in the H1NMR spectra of diastereomeric imines as was the % ee for 2-methyl1-tetralone when compared to that for its diastereomeric imine. Thus, hydrolytic conditions were the major cause of the low ee's observed for entries 11 and 12 in Table I. We next turned our attention to asymmetric dialkylation of cyclohexanone in an effort to reach chiral, nonracemic 2,6-dialkyl derivatives. When (S)-2 was metalated with LDA in THF, followed by addition of methyl iodide (-78 "C), the initial diastereomeric product (12) was assumed to be the syn-axial product on the basis of work of Fraser.14 Addition of a second equivalent of LDA, followed by heating the solution to reflux (2 h), gave the isomerized imine (anti-equatorial) 13 which underwent efficient deprotonation to the metalloenamine 14. Subsequent addition of methyl iodide (-78 "C) furnished the dialkylated imine 15 which was hydrolyzed in the acetic acid sodium acetate buffer to (ZS,6S)-(+)-2,6-dimethylcyclohexanone(16a) in 85% ee.15 It is important to note that without heating 12 in the presence of LDA prior to the addition of methyl iodide, considerable monoalkylated ketone (2040%) is recovered. In a similar fashion, n-propyl iodide was utilized to dialkylate the chiral imine of cyclohexanone and gave (2S,6S)-(+)-2,6-bis(n-propyl)cyclohexanone (16b) in 91% yield.I6 Again, if the monoalkylated imine (13) Jaouen and Meyer (Table I, footnote g) report that the (-)-2methyl-1-indanone possesses the R configuration, in contrast to our expectation that the (-)-enantiomer should be S based on all the examples in this and the following paper. Furthermore, Kagan et al. (Tetrahedron 1971, 27, 4737) report that (+)-2-methyl-1-tetralone possesses the S configuration in contrast to Jaouen and Meyer and our own assignment. This matter therefore remains unresolved and could represent a deviation from the predicted absolute configurations reported for all other examples in Table I. (14) Fraser, R. R.; Banville, J.; Dhawan, K. L. J . Am. Chem. SOC.1978, loo, 7999. (15) Beard, C.; Djerassi, C.; Snider, J.; S i p s , F.; Tichy, M. Tetrahedron 1963, 19 report [a]3~ +130°. Chiral LISR studies confirmed the enantiomeric purity; see Experimental Section.

OMe

(ss) -13 (ant i-eq)

\

ss, 93i31

(reflux, 2h)

H

R

j &

LDA-THF

/Ph

Me

OMe

+

1 I.?

OMe

x a ,

R=Me, lalD +10W

IBb,

R:n-Pr,

lalD +59.B0

12 was not heated with LDA prior to the second propyl iodide addition, 67% of 2-propylcyclohexanone was isolated along with only 33% of the 2,6-di-n-propyl derivative. The previously unreported enantiomeric purity of 16b could not be assessed by any usual means (LISR, HPLC, or N M R of diastereomers) although its CD spectrum exhibited +3977 which is comparable both in sign and intensity to 2,6-dimethylcyclohexanone16a ( +3868). This suggests that its enantiomeric purity and absolute configuration were closely related to those of 16a. This double alkylation may be seen as initially giving syn-axial alkylation to 12 (via syn-axial) deprotonation of 2 in accordance with the observations by Fraser.14 Isomerization of 12 to 13 and ultimately to the metalloenamine 14 probably occurs via heat followed by metalation. Alternatively, it could take place by metalation of 12 followed by rapid isomerization. Choice between these alternate pathways is not possible at this time. However, one fact remains clear. The metalation of 12 appears to be very slow unless heat is applied in the presence of the base. This apparent stereoelectronic effect toward deprotonation is one which has received considerable attention in other weak organic acids and suggests that this phenomenon may become potentially very important for stereochemical and regiochemical control in synthesis." This intriguing behavior prompted another experiment which was designed to provide further insight into this stereoelectronic effect. Metalation of 2 (LDA, -20 OC, THF), followed by isopropyl iodide gave the 2-isopropyl imine 17 which was treated with LDA at -20 "C, or at reflux, followed by addition of methyl iodide. To our surprise only 18% yield of 2-methyl-6-isopropylcyclohexanone (19) was obtained along with 80% of 2isopropylcyclohexanone (18). The enantiomeric purity of the latter was only 5-10% due to considerable racemization during the hydrolysis of the imine. However, the significant result rested with the ratio of the mono- (18) to the dialkylated (19) product. (1 6) Only the 2-axial-6-equatorial cyclohexanones are chiral and would exhibit optical activity. The 2,6-diequatorial derivatives are meso compounds. (17) Meyers, A. I.; Campbell, A. L.; Abatjoglou, A. G.; Eliel, E. L. Tetrahedron Lett. 1979, 4159. Abatjoglou, A. G.; Eliel, E. L.; Kupyer, L. F. J . Am. Chem. SOC.1977, 99, 8262. Wolf, S. Ace. Chem. Res. 1972, 5 , 102. Lehn, J. M.; Wipff, G. J . Am. Chem. Soc. 1976, 98, 7498. Streitweiser, A.; Williams, J. E. Ibid. 1975, 97, 191.

J . Am. Chem. Soc., Vol. 103, No. 11, 1981 3085

Enantioselective Alkylation of Ketones Scheme III

- /ja R :: +N/

20 -

6-

21 (70%) +

a) LDA- M e 1

H@

b)

'

I I .(30%)

T

18,

R

=

H

19, -

R

=

Me (80%)

(20%)

4000

-2

L DA -1-Pr I

;I"p a)

18, (80%)

LDA-Me1

19, (18%)

b)

17 -

-2 -

LDA-Me1

a) LDA-L-PrI

18, (N2%)

b)

2, (95%)

H@

Me

12 -

On the other hand, reversing the order of sequential metalationalkylation on 2-metalation and introduction of methyl iodide to 12 followed by metalation (reflux)-isopropyl iodide gave the dialkylated ketone 19 in 95% yield with only a trace of 2methylcyclohexanone recovered. Once again, the % ee of the ketone 19 was low (20-30%) due to racemization during hydrolysis. Although the latter result was not unexpected, the former [ 17 18 (SO%)] was indeed surprising. The 2-isopropyl imine 17 was then isolated and its 13CNMR spectrum examined which revealed the presence of a single syn-axial isomer.2s Signals for the isopropyl-substituted carbon (C-2) at 6 45.0 and the d-carbon at 6 37.0 are in agreement with previous assignment^.'^ Heating 17 more drastically (80-100 "C) resulted in a major change in its 13CN M R spectrum and indicated that 17 had isomerized into a 70:30 mixture of 21 and 17 (Scheme 111). Metalation of this mixture (LDA, -20 "C, MeI) gave, in contrast to reaction of pure 17, a 2030 ratio of 18 and 19, respectively. Furthermore, when racemic 2-isopropylcyclohexanone (20) was converted to its imine by using excess (S)-1, the product was virtually identical by 13C N M R with that obtained after heating imine 17. Metalation of this mixture from (f)-20 gave 18 and 19 in the same ratio (20:80). One may conclude from this behavior that the isopropyl group enters in a syn-axial manner (17) but is simply too bulky to isomerize under mild conditions (refluxing THF) to the antiequatorial isomer to allow metalation. Alternatively, metalation may be inhibited in 17 due to the presence of the axial isopropyl group. It is also of interest that the isopropyl imine formed from 2-isopropylcyclohexanone (20) gives rise to a thermodynamic mixture of imines by virtue of its method of formation which is indeed quite different than the kinetically controlled metalation-alkylation of cyclohexanone imine, 2.

-

Experimental Section1* (S)-(-)-2-Amino-l-methyoxy-3-phenylpropane(1). Reduction of (S)-Phenylalanine. A modified procedure, previously reportedi9Pand (18) Microanalyses were performed by Midwest Microlabs, Indianapolis, Ind. Optical rotations were taken on a JASCO DIP-180or a Perkin-Elmer 241 automatic polarimeter. 'H NMR were taken on a Varian 360A or T-60 instrument and I3C NMR were performed with a JEOL FX-100instrument operating at 25.05 MHz. Gas chromatography was performed on a Hewlett-Packard 5750 FID chromatograph using SE-30, 6 ft X '/* in. column.

Table 111. Chiral Imines from (9-1 and Cyclic Ketones VC=N

ketonea

Kugelrohr, "C (pressure, torr)

(film), [CY]~~~(CHC cm-' ~,)

cycloheptanone 90-100 (0.1) -42.91' (C 8.6) 1640 134-138 (0.025) -53.62" (c 6.7) 1645 cyclooctanone 1-tetralone 160-164 (0.15) -207.9' (C 6.64) 1630 a These ketones re uired 1-5% trifluoroacetic acid as a catalyst to form the imines. The ' H NMR spectra were recorded but were not of any unusual nature.

'b

based on Yamada'slgb technique, was employed. (S)-Phenylalanine (183.4 g, 1.11 mol) was suspended in 1 L of methanol (dried over Na2S0,). The solution was cooled to 0 "C (ice-water bath) and 200 mL (2.74mol) of thionyl chloride was added dropwise over ca. 1 h. the solution was allowed to warm to room temperature and stirred overnight. Concentration gave a yellow solid which was stirred with 400 mL of Et20, filtered, and dried under high vacuum (to remove excess SOC12), to produce 236.5 g (101%) of the white solid hydrochloride salt of the methyl ester. The ester (0.497mol), without further purification, was dissolved in 500 mL of 50% aqueous EtOH. This was slowly added to a cooled (0 "C) solution of NaBH, (89.8g, 2.38 mol) in 1500 mL of 50% aqueous EtOH over Ca. 1 h (gas evolution and insoluble liquid layer at bottom). The solution was stirred at room temperature for 4 h, then refluxed under N2 for 6 h, and allowed to stir at room temperature overnight. The top layer was decanted and the ethanol removed in vacuo until a white solid appeared. The lower insoluble layer was extracted (3 X 150 mL) with EtOH and concentrated. The two concentrated aqueous layers were extracted (5 X 100 mL) with EtOAc. The combined EtOAc layers were dried over Na2S04 and concentrated to yield a white solid. This was recrystallized in EtOAc (255 mL) and hexane (100 mL) to give 58.3 g (78%) of the desired alcohol: 'H NMR (CDC13) 6 1.53 (s, 2 H), 2.78 (m, 2 H), 3.23 (d, 2 H), 3.45-3.70(m, 1 H), 4.20(s, 1 H), 7.15 (s, 5 H); mp 89-91 "C; ["ID -24.2" ( C 1.5, EtOH). Methylation with KH-Me1 to (S)-l. This was performed as previously described;19a [a]25D -13.8' (c 5.6,benzene). The product, (S)-1, is prone to carbon dioxide adsorption to form the carbonate and was +19.5" (c 2.5,EtOH) or kept either stored as its hydrochloridei9([a]*$,

(19)(a) Meyers, A. I.; Poindexter, G. S.; Brich, Z . J. Org. Chem. 1978, 43, 892. (b) Seki, H.; Koga, K.; Matsuo, S.; Ohki, I.; Matsuo, I.; Yamada, S. Chem. Pharm. Bull. 1965,13, 995.

3086 J . Am. Chem. SOC..Vol. 103, No. 11, 1981

Meyers et al.

Chemical Shifts of Acetals 9 and Diastereomers at C-2 and C d

Table IV.

c-1

c-2

c-3a

C-4

c-5

Cd

othersb

Me

109.7

39.6 40.0

24.9

32.2

23.1

31.2 31.7

Et

109.8

46.1 47.4

24.1 25.2

28.1

23.6

31.2 37.9

n-Pr

110.0=

44.5 45.2

24.1 24.6

28.6

23.6

36.5 37.1

PhCH,

109.5

46.9 41.6

24.8

28.1

23.6

36.6 31.4

2CH,, 14.4 CB,79.6 C-7. 71.6 (2-9; 16.2; C-10, 17.8 CH,, 20.9;CH3, 11.9 CB, 79.3 c-I,77.3 C-9, 16.2; C-10, 17.9 CH,, 20.5; CH,, 30.2 CH,, 14.4 C-8, 78.9 C-I, 17.5 C-9,16.1; C-10, 17.8 CH&, 34.8 AI, 121,128,125,141 C-8,79.2 C-I, 11.5 C-9, 16.3; C-10, 17.9

R

a Diastereomers were also discernible a t C-3 but were not generally useful due to small AA6 values in some cases. Spectra of these comDiastereomers could also be seen at C-1, but the AAS were too small to be useful and pounds may be found with supplementary material. these peaks did not separate in all cases.

to be 98 f 2% pure. Specific rotations and CD data are given in Table tightly sealed under argon or nitrogen immediately after distillation, bp I. 57-59 OC (0.02 torr). Cyc~hexanowImine of ( ~ ) - ( - ) - 2 - A ~ l - ~ ~ x y - ~ p ~ y l p Physical r o ~ Data for 2-Alkyl Ketones. (a) (R)-(+)-2-Acetonylcyclohexanow:2' bulb-to-bulb distillation 50-60 OC (0.05 torr); IR (film) (2). In a system containing a Dean-Stark trap arranged for azeotropic 1710 cm-I; NMR (CDCI,) 6 2.1 (s, 3). VPC shows 99% purity. removal of water, the methoxy amine 1 (9.3 g, 56.4 mmol) and cyclo(b) (R)-( +)-24 2-Methoxyethyl)cyclobexanone: bulb-to-bulb distilhexanone (6.5 g, 66.4 mmol) were dissolved in 100 mL of benzene or lation 80-90 OC (0.05 torr); IR (film) 1705 cm-'; NMR (CDCI,) 6 toluene and heated to reflux until the theoretical amount of water was 1.1-2.6 (m, 1l), 3.20 (s, 3), 3.25 (t, 2). VPC shows 98% purity. Anal. collected in the trap. Removal of the solvent and distillation of the oily Calcd for C9HI6O2:C, 69.23; H, 10.26. Found: C, 68.81; H, 10.27. residue gave 13.0 g (93%) of a clear viscous oil: Kugelrohr 95-100 OC (c) (S)-( +)-2-Methyl~ycloheptanone:~~ bulb-to-bulb distillation 50 (0.5 torr); [ a ]-46.2' ~ (C 6.7, MeOH). OC (8 torr); IR (film) 1695 cm-'. VPC shows 99 1% purity. In a similar fashion the other cyclic ketones were converted to their (a) (R)-(+)-2-(2-Methoxyethyl)cycloheptanone: bulb-to-bulb disimines by using 1 (Table 111). tillation 80-90 OC (0.05 torr); IR (film) 1700 cm-I, NMR (CDCIJ 6 Metalation and Alkylation of Cyclohexanone Imines. Asymmetric 1.20-2.50 (m, 13), 3.30 (s, 3), 3.35 (m, 2). VPC analysis shows 99% Synthesis of 2-Alkylcyclohexanones 5. An oven-dried 50-mL flask purity. Anal. Calcd for CtnHlnO,: equipped with a magnetic stir bar, a pressure-equalizedaddition funnel, .- .- - C, 70.55; H, 10.66. Found: C, i m i ; H, 10.71. and a rubber serum cap was charged with 20 mL of anhydrous tetrale), ( S b ( + ~ - 2 - M e t h v l ~ v c bulb-to-bulb l ~ : ~ ~ distillation 50 OC hydrofuran under a nitrogen atmosphere. Freshly distilled diisopropyl(10 torr); IR ;film) 17iO>m-'; NMR (CDC13) 6 1.00 (d, 3), 1.10-2.00 amine (1.47 mL, 10.5 mmol) was added via syringe and the solution (m, 10). VPC shows 3% unalkylated cyclooctanone, [ a ] D +7.76, corcooled to 0 OC. Butyllithium (4.4 mL of a 2.4 M solution in hexane) was rected for pure ketone, [aID+8.07. added and the solution stirred at 0 OC for 15 min and then cooled to -30 (f) (S)-(-)-2-Methyl-l-tetra10ne:~~ bulb-to-bulb distillation 85-100 in 10 mL of THF OC. The chiral imine of the various ketones (10 "01) OC (0.1 torr); IR (film) 1690 cm-'; NMR (CDCI,) 6 1.27 (d, 3), 2.17 (dry)was added over 15 min and allowed to stir for 1-1.5 h. The solution (m, 2), 2.75 (m,l), 3.00 (m, 2), 7.40 (m,3), 8.10 (m,1). VPC showed was then cooled to -78 OC and the alkyl halide (10.5 mmol) was added 99% purity. in a solution of THF (3-4 volumes) over a period of 1 h and the mixture (8) (R)-(-)-2-Methyl-l-indan011e:~~ bulb-to-bulb distillation 35-45 allowed to stir at -78 OC for 1.5 h. The total volume of THF was such as to make the final concentration of the imine -0.25 M. The light OC (0.005 torr); IR (film) 1710 cm-I; NMR (CDCl,) 6 1.20 (d, 3), 2.30-2.80 (m, 2), 3.10-3.60 (m,l), 7.10-7.80 (m. 4). VPC showed 99% yellow cold solution was poured into 100 mL of saturated salt solution purity. and extracted with ether (3X). The combined ether extracts were washed Recovery of 1. The acidic aqueous layer from above was neutralized with brine, then dried (Na2S04), and concentrated in vacuo, and the with solid KOH to pH 14 and then saturated with sodium chloride. This amber oil was subjected to hydrolysis immediately. Delay of the hyaqueous solution was extracted four times with ether, and the combined drolysis step (e.g., overnight) results in 10-15% racemization. A buffer ether layers were washed with brine. Drying (K2C0,) and concentration solutionmcomprised of 3.30 g of sodium acetate, 7.5 mL of acetic acid, gave an oil which was distilled: bp 57-59 OC (0.02 torr); [ a ] D -13.75' and 35 mL of water was added to the crude imine in 50 mL of pentane (c 5.8, benzene). The recovery was -85%. If the rotation of the distilled and shaken for 30 min and the aqueous layer removed and saved to methoxyamine 1 was lower than 13.40 "C, it was purified via its hyrecover 1 (see below). The aqueous layer was washed with pentane, and drochloride. A solution of 1 in ice-cold ether was treated with dry HCl the pentane solutions were combined and washed with 1 N HCl (to by bubbling through a fritted disk. The amine hydrochloride was colremove any unhydrolyzed imine), water, 5% sodium bicarbonate, water, lected by filtration and recrystallized from ethanol: mp 151-152 OC; and brine. The pentane solution was then dried (Na2S04), filtered, concentrated, and bulb-to-bulb distilled to yield the 2-substituted ketone. Alternatively,hydrolysis could be carried out with equally satisfactory (21) Islam. A. M.; Raphael, R. A. J. Chem. Soc. 1952,4086. results using 3 mL of saturated (&)-tartaric acid solution and 2 mL of (22) Gutsche, C. D.; Cheng, C. T. J. Am. Chem. Soc. 1962, 84, 2263. pentane/mmol of imine obtained after solvent removal. The 2-alkyl These authors report Q-(+)-2-methylcycloheptanone [a]D +29O as 39% ee ketones thus obtained were examined by gas chromatography and shown which would correspond to optically pure material having [a]D 73.8'. on the basis of this value, 2-methylcycloheptanonewould be 97% ee. The 85% ee reported in Table I was determined by LISR technique (see footnote A. (23) Cope, A. C.; Ciganek, E.; Lazar,J. J. Am.Chem. Soc. 1%2,84,2591. (20) Kieczykowski, G. R.; Schlcssinger,R. H.; Salsky, R. B. Terruhedron (24) Jaouen, G.; Meyer, A. J. Am. SOC.1975, 97,4667. Lett. 1976, 597.

*

. .-, .

-

Enantioselective Alkylation of Ketones ['Y]D +19.50" (c 2.5, ethanol). Release of the amine 1 from its hydrochloride was accomplished by dissolving it in water containing 15% sodium chloride, adding solid KOH to pH 14, and extracting with ether (4X). Washing the ether solution with water and then with brine, and drying (K2C03)gave, upon concentration, the amine ['Y]D -13.75' in purity sufficient for imine formation. Cyclic Acetals 9 of Cyclohexanone and (ZR,3R)-2,3-ButanedioI. The following procedure was used to prepare the cyclic acetals of 2-methyk 2-ethyl-, 2-n-propyl-, and 2-benzylcyclohexanones. The ketone (200 mg) and the (-)-diol (2 equiv) in 15 mL of methylene chloride containing 3-5 mg of ptoluenesulfonic acid were heated at reflux for 1-20 h depending on the 2-alkyl substituent. The reaction was followed by VPC (6 ft X in., SE-30 on 60-80 mesh Chromsorb W, 90-150 OC 10 "C/min), and heating continued until all of the ketone was gone. This avoided any kinetic resolution of the ketones. After being cooled, the methylene chloride solution was washed with water and then saturated bicarbonate, dried (Na2S04), and concentrated in vacuo. The resulting products (VPC purity, 97-99%) were obtained in 85-95% yield, and their I3C NMR spectra were examined as CDCI3 solutions at 25.05 MHz (Table IV). (2S,6S)-(+)-2,6-Dimethylcyclohexanone(16a). A solution of 1.23 g (5.0 mmol) of cyclohexanone imine 2 in 5 mL of THF was added to a solution of LDA (5.20 mmol) in 10 mL of THF at -20 OC over 15 min. The resulting yellow solution was stirred for 1 h under a nitrogen atmosphere and then cooled to -78 "C. To this was added 0.78 g (5.5 mmol) of methyl iodide in 5 mL of THF, and the clear mixture was stirred for 2 h at -78 "C and allowed to warm to ambient. This solution was transferred, via nitrogen pressure, through a cannula (double-edged syringe needle) into an addition funnel (pressure equalized attached to a three-necked flask containing 5.25 mmol of LDA in 10 mL of THF and equipped with both a nitrogen inlet tube, a reflux condenser, and a magnetic stir bar. The LDA solution was cooled to -20 "C and the imine solution added from the addition funnel over 15 min. The yellow solution was stirred for 1 h at -20 OC and then warmed to ambient and heated to reflux for 45 A n . The somewhat darkened solution was cooled to -78 "C and a solution of 0.80 g (5.7 mmol) of methyl iodide in 5 mL of THF added, and the mixture was stirred at -78 OC for 2.5 h and warmed to ambient after which it was poured into 150 mL of saturated brine and worked up and hydrolyzed as described for the monoalkylation of cyclohexanones. Kugelrohr distillation gave 0.48 g (77%) of a clear colorless liquid: bp (bulb-to-bulb) 80-90 "C (9 torr); ['Y]"D +88.8" (c 7.8, MeOH) containing 18% meso-(2R,6S)-2,6-dimethylcyclohexanone. Correcting for the mesocontaminant gave [aID+108O which was 83% ee based on literature vahe15 of [ a ] D +130°; CD (MeOH) at 292, [e] = +3868 (lo-) M). Verification of the meso compound was achieved by use of Resolve-A1 (Aldrich) an achiral shift reagent (20 mg/0.025 g of cyclohexanone in 0.4 mL of CC14) which showed two doublets in 82:18 ratio. The % ee was also verified by use of europium shift reagent (footnotef, Table I): 18-20 mg/0.3 g of ketone in 0.4 mL of CC14which gave integration of the doublets at 85 & 2%. (2S,6S)-(+)-2,dDipropylcyclohexanone (165). In a manner similar to that above, 5 mmol of cyclohexanone imine 2 was treated with LDA and n-propyl iodide in sequential fashion and heating the monopropylated imine with LDA prior to addition of the second equivalent of n-propyl iodide. Workup and hydrolysis gave 165 in 91% yield: bp (bulb-to-bulb) 80-90 "c (0.02 torr), [ C ~ ] ~ ' D+59.6" (c 3.1, MeOH). VPC (UCW-98 on chromsorb W, 60-80 mesh) showed 2% mono-n-propylcyclohexanone: mass spectrum, ( m / e ) 182 (M+), 140, 111, 98, 91, 83, 77, 67, 55, 41; IR (film) 1715 cm-I; 'H NMR (CC14)6 2 . 6 0 . 6 0 (m,22); CD (MeOH) M). Anal. Calcd. for C1zH2z0:C, 79.20; at 292, [e] = +3977 H, 12.10. Found: C, 79.42; H, 12.17.

J. Am. Chem. SOC.,Vol. 103, No. 11, 1981 3081 (R)-(+)-2-Isopropylcyclohexanone(18). To 4.4 mmol of LDA in 15 mL of THF at -20 "C contained in a flame-dried flask fitted with additional funnel, N 2 inlet, magnetic stirrer, and septum was added 0.98 g (4.0 mmol) of the imine 2 in 10 mL of THF. The resulting solution was stirred at -20 "C to 0 "C for 30 min and then cooled to -78 "C. The lithioenamine was alkylated with 1.02 g (6.0 mmol) of 2-iodopropane in 5 mL of THF. The resulting solution was allowed to warm to ambient over a 3-h period (NOTE: warming to ambient is necessary in this particular alkylation to obtain complete alkylation) and quenched with brine. The organic layer was separated and extracted 3X with ether, and the combined organics were washed with brine. The solvents were removed in vacuo, and the residue was examined by 13C NMR. NMR analysisz5indicated the syn isomer to be present in >95% diastereomeric purity. Hydrolysis of the imine with NaOAc/HOAc/pentane buffer for 13 h, pentane extraction, and subsequent Kugelrohr distillation (bp 50-80 "C (0.05 torr)) produced 0.46 g (83% yield) of the desired product 18. VPC analysis (1/4 X 6 ft, SE-30, 100-175 "C) indicated the presence of only one peak. ['YIz5D +3.7" (c 6.13, CHC13). The enantiomeric excess was determined to be 5% of the R enantiomer on the basis of the maximum reported rotation4 of ['YIz3D -75" (c 1.56, MeOH) for the S enantiomer. 2-Methyl-disopropylcyclohexanone(19). The imine (5.0 mmol, 1.22 g) from cyclohexanone and 2-amino-1-methoxy-3-phenylpropanein 5 mL of THF was added to 5.25 mmol of LDA at -20 OC in 5 mL of THF contained in a 50-mL flame-dried flask. The resulting solution was allowed to warm to 0 "C and stirred for 1 h before cooling to -78 "C. Methyl iodide (10 mmol) in 5 mL of THF was added dropwise and the solution allowed to warm to 0 OC. Another 6.0 mmol of LDA in 10 mL of THF was added to the reaction and the resulting solution heated to reflux for 45 min. The solution was then cooled to -78 OC and 2-iodopropane (7.0 mmol) in 5 mL of THF was added dropwise. The resulting soluton was allowed to slowly warm to room temperature and quenched with saturated NaC1. The organic layer was separated and the aqueous layer extracted three times with Et20. The organics were combined and washed twice with brine. The solvent was removed on a rotary evaporator and the residue hydrolyzed with NaOAc/HOAc/pentane buffer for 13 h. The organic layer was separated and the aqueous layer extracted with pentane. The organics were combined and washed with H 2 0 and dried over Na2S04, and the solvent was removed on a rotary evaporator. Kugelrohr distillation of the residue (80-100 OC (0.05 mm)) gave 0.61 g (80% yield) of the desired product. VPC analysis (10% SE-30, 100-160 "C) indicated 3% of 2-isopropylcyclohexanone and 97% of 2-methyl-6-isopropylcyclohexanone.IH NMR (CDCIJ 6 1.4-2.5 (8 H, overlapping signals), 0.6-1.3 (10 H, doublet at 6 1.0, J = Hz, doublet Of doublets at 6 0.9, J = 4 HZ);[ a I Z 5 +21.96" ~ (C 4.69, CHCI,).

Acknowledgment. We are grateful to the National Science Foundation (Grant CHE-7905580) for financial support of this work. The 13CNMR spectra were provided by the Colorado State Univeristy Regional NMR Center funded by the National Science Foundation (Grant CHE-78- 18 58 1) . Supplementary Material Available: The spectra for the 13C N M R determination of enantiomeric excess of acetals 9 (12 pages). Ordering information is given on any current masthead page. ( 2 5 ) "C NMR spectra of 17,a mixture of 17 and 20 after heating, and deuterated compounds, as well as 13CNMR spectra for actals 9, are included as supplementary material.